They act like reasoning about safety of automated systems is brand new area that they need to develop and disrupt.
Sure, heavy reliance on AI makes some traditional forms of validation more tricky, but BS metrics and test in prod is so reckless, that will end up only hurting the industry. When there will be a big tragedy, because of reckless approach, government will be forced to drop heavy hammer of regulations and make it much harder to develop self driving tech.